Author + information
- Patricia P. Chang, MD, MHS∗ ()
- Division of Cardiology, Department of Medicine, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
- ↵∗Address for correspondence:
Dr. Patricia P. Chang, Division of Cardiology, Department of Medicine, The University of North Carolina at Chapel Hill, CB 7075, 160 Dental Circle, 6th Floor Burnett-Womack, Chapel Hill, North Carolina 27599-7075.
Trends in hospitalizations for acute heart failure (HF) may be increasing, primarily driven by acute HF with preserved ejection fraction (HFpEF) (1). However, HF may not be the main reason for hospitalization in the last year of life. One-year mortality remains high (approximately 30%), both in the United States (2) and Denmark (3), regardless of ejection fraction (EF) (1). The overall cost of HF continues to rise, mostly related to direct medical costs (2). The greatest burden of any chronic disease is manifested in the intensity and cost of medical care, as well as resource use in the last year of life. Whether this is a global pattern for patients with HF is less clear.
In this issue of JACC: Heart Failure, Madelaire et al. (4) reported that most Danish patients with HF had at least 1 hospitalization during the last year of life, which was mostly for noncardiovascular causes (4). Using the Danish National Patient Registry, patients were included in the study if: 1) they received a first inpatient or outpatient HF diagnosis from 2000 to 2015 based on 3 specific International Classification of Disease-10 (ICD-10) diagnosis codes (I50: HF, I42: cardiomyopathy, or J81.9: pulmonary edema) as either a primary or secondary diagnosis; 2) they were started on therapy with an angiotensin-converting enzyme inhibitor or angiotensin receptor blocker (ACEI/ARB) and beta-blocker (BB) within 4 months after the HF diagnosis; and 3) they died by the end of 2016. Because EF data were not available to subclassify cases as HF with reduced EF (HFrEF) or HFpEF, the inclusion criteria of dual neurohormonal therapy aimed to increase (or bias) the proportion of HFrEF cases.
Of the 32,157 Danish decedents with HF, 67% were 75 years and older (median age 81 years), 61% were men, and 23% were originally diagnosed as outpatients. Median duration of HF was 4.2 years at the start of the study, which was consistent with a median survival of approximately 5 years as previously noted in this Danish population registry (3) and in the United States (2). At 1 year before death, only 80% and 84% of the sample were taking ACEIs/ARBs or BBs, respectively, which suggested some intolerance to these therapies, but only 84% were on loop diuretic agents, which suggested that HF was not symptomatic for a sizeable minority. Interestingly, up to one-third of patients were on sedatives, antidepressants, anxiolytics, and/or opioids, which reflected a relatively high prevalence of chronic mental health comorbidities or pain issues.
Not surprisingly, most patients (83%) were hospitalized during the last year of life with a median of 2 hospitalizations and a relatively long median length stay of 2 weeks. The 17% of patients who were not hospitalized were older, more often lived in a nursing home, or had dementia. Among those with hospitalizations during the last year, most (64%) were for noncardiovascular causes, which was several-fold higher than the 12% due to HF and 18% due to other cardiovascular (CV) causes. The frequency of hospitalization increased toward death, especially during the last couple of months of life, with non-CV hospitalizations comprising the vast majority throughout the last year, regardless of age, sex, and comorbidity burden (as estimated by the Charlson Comorbidity Index). However, those hospitalized for HF or other CV causes were younger, had more CV comorbidities, and had more frequent use of therapies typical for patients with HFrEF (ACEIs/ARBs, BBs, implantable cardioverter-defibrillators, cardiac resynchronization therapy). Notably, diuretic dosing was intensified after hospital discharge in 18% of all hospitalizations (including 15% of the non-CV hospitalizations), which suggested contribution of worsening HF symptoms in a minority of cases. Although HF cases could not be subclassified by left ventricular EF, the important point remains: most hospitalizations during last year of life were for non-CV causes for all ages, for both sexes, and throughout the time period.
This study also provided important insights about trends of HF incidence, therapies, and outcomes. During 15 years, more HF cases were diagnosed (or more HF-related ICD-10 codes were noted) as outpatients, more defibrillators and/or cardiac resynchronization therapy devices were implanted, non-CV hospitalizations increased, and slightly lower proportions of patients died in the hospital.
Although this registry study was impressively large, the main limitation was that the HF case definition was based on ICD-10 codes in any position (primary or secondary). As with many other studies using administrative data (e.g., Medicare claims and the National Inpatient Sample in the United States), the main issue with defining HF cases in this manner is the potential for misclassification, because ICD codes are used primarily for billing purposes and may be listed for differential diagnosis evaluation. Therefore, HF cases identified solely by ICD codes are prone to overestimation. However, misclassification is less likely when ICD codes are in the primary position, as was used in their study for classifying subsequent HF hospitalizations from other causes for hospitalization. For example, in the setting of the ARIC (Atherosclerosis Risk In Communities) study’s community surveillance of hospitalized HF events, HF-specific ICD codes (ICD-9 code 428, similar to ICD-10 code I50) in the first position had a 93% positive predictive value and 95% specificity for a validated acute HF hospitalization (5). Madelaire et al. (4) could have performed an additional sensitivity analysis to limit case definition to the specified ICD-10 codes in the primary position, which could have reduced potential misclassification of patients being ruled out for HF.
This study remains important because it provides a European, and thus, a more global perspective about the last year of life and also confirms similar results from U.S. patient populations. In the Rochester Epidemiology Project cohort study that occurred during the same time (2003 to 2011), most patients (81.5%) in this southeastern Minnesota cohort were also hospitalized at least once during the last year of life (median 2 hospitalizations per person) (6). Hospitalizations and days hospitalized increased in the months before death, particularly in the last few months. Although the patient sample in this cohort study was smaller (n = 683) and more selected (patients provided consent), HF cases were confirmed using Framingham criteria and subclassified using echocardiographic data, and subsequent hospitalization and outpatient visits were ascertained using ICD-9 codes from administrative data. Among these Minnesotan patients with HF, 53% had HFpEF (left ventricular EF ≥50%) and mean age of death was 82 years, but more than one-half (53%) were women. Older age and dementia were also associated with lower use of hospitalization and outpatient visits, whereas patients who were married, had more comorbidities, or resided in a skilled nursing facility had higher use. Notably, patients with HFpEF had high rates of non-CV hospitalization.
With our greater appreciation of end-of-life care, palliative care and hospice use should be encouraged when patients are more frequently hospitalized. Over the 8 years in the study by Dunlay et al. (6), rates of hospitalization and outpatient visits decreased, whereas palliative care consults and enrollment in hospice increased; these trends were similar in patients with HFpEF and HFrEF. Compared with the Danish sample in the study by Madelaire et al. (4), of whom 45% died in hospital (4), fewer Minnesotans (28%) died as inpatients (6). Similar trends have been seen among Medicare beneficiaries. Thus, we should be encouraged to engage our palliative care colleagues when we see our older adult patients with HF hospitalized at least twice within a 6- to 12-month period. Alternatively or in addition, for those patients with HFrEF who may be eligible for advanced HF therapies (e.g., left ventricular assist device support or heart transplantation), we should consider these surgical therapies earlier rather than later.
Care of the patient in the last year of life is expensive because we spend more on the sick, both on those who ultimately die and on those who recover or improve. Assuming much of that cost is hospitalization, this study by Madelaire et al. (4) emphasizes that the greatest burden of hospitalizations for patients with HF is for non-CV disease care management. Although we must pay attention to the heart, we should also focus on the whole patient and use palliative care and/or hospice when appropriate, which ultimately can decrease costs and medical care use. In the end, it is more of an art than science in how to optimize the last year of life.
↵∗ Editorials published in the JACC: Heart Failure reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Dr. Chang has reported that she has no relationships relevant to the contents of this paper to disclose.
- 2019 American College of Cardiology Foundation
- Chang P.P.,
- Wruck L.M.,
- Shahar E.,
- et al.
- Benjamin E.J.,
- Muntner P.,
- Alonso A.,
- et al.
- Schmidt M.,
- Ulrichsen S.P.,
- Pedersen L.,
- et al.
- Madelaire C.,
- Gustafsson F.,
- Kristensen S.L.,
- et al.
- Rosamond W.D.,
- Chang P.P.,
- Baggett C.,
- et al.
- Dunlay S.M.,
- Redfield M.M.,
- Jiang R.,
- et al.